AI Units
A consumption-based metering unit introduced by SAP for AI services, used to bill AI functionality—including Joule scenarios, ABAP AI features, and agent orchestration calls—within RISE and GROW contracts. The number of units consumed varies by operation complexity.
Timeline (last 90 days)
AI Units are the metering unit SAP uses for its AI services (such as Joule and model calls in AI Core), similar to Tokens or API call counts in cloud services; SAP bills customers based on consumption.
Architects need to set an AI Units budget for each agent, monitor consumption curves, design degradation strategies, and incorporate cost routing decisions into launch reviews, just as they plan CPU and storage capacity.
For SAP architects, the article proposes a 12-month skills plan: in the first quarter, learn the MCP threat model and build a non-production agent with read-only tools; in the second quarter, add observability and write a permission matrix; in the third quarter, practice cost routing patterns and document AI Units forecasts; in the fourth quarter, lead a governed production use case end to end.
Dev/production separation is the single most effective cost control measure; restricting AI features to development and quality systems alone typically reduces consumption by 60–80%
Usage forecasting methodology recommends reserving a 20–30% buffer for experimentation and unexpected consumption
SAP will officially commercialize AI Units billing in 2026
If a log analysis MCP agent consumes 500 SAP AI Units per run and runs 20 times per month, switching to a local deployment can save 10,000 units
The sample monthly total forecast for a 200-user system is approximately 371,250 AI Units
A simple syntax check may consume only 1–5 AI Units, while a complex ABAP AI refactoring may consume 50–200 units
In agent orchestration scenarios, each tool call is an independent round-trip request at the protocol level and counts toward AI Units consumption, making cost volatility far higher than when using Joule directly
All Facts (10)
For SAP architects, the article proposes a 12-month skills plan: in the first quarter, learn the MCP threat model and build a non-production agent with read-only tools; in the second quarter, add observability and write a permission matrix; in the third quarter, practice cost routing patterns and document AI Units forecasts; in the fourth quarter, lead a governed production use case end to end.
50%UnverifiedAI Units are the metering unit SAP uses for its AI services (such as Joule and model calls in AI Core), similar to Tokens or API call counts in cloud services; SAP bills customers based on consumption.
50%UnverifiedArchitects need to set an AI Units budget for each agent, monitor consumption curves, design degradation strategies, and incorporate cost routing decisions into launch reviews, just as they plan CPU and storage capacity.
50%UnverifiedDev/production separation is the single most effective cost control measure; restricting AI features to development and quality systems alone typically reduces consumption by 60–80%
50%UnverifiedUsage forecasting methodology recommends reserving a 20–30% buffer for experimentation and unexpected consumption
50%UnverifiedIf a log analysis MCP agent consumes 500 SAP AI Units per run and runs 20 times per month, switching to a local deployment can save 10,000 units
50%UnverifiedThe sample monthly total forecast for a 200-user system is approximately 371,250 AI Units
50%UnverifiedA simple syntax check may consume only 1–5 AI Units, while a complex ABAP AI refactoring may consume 50–200 units
50%UnverifiedIn agent orchestration scenarios, each tool call is an independent round-trip request at the protocol level and counts toward AI Units consumption, making cost volatility far higher than when using Joule directly
50%UnverifiedSAP will officially commercialize AI Units billing in 2026
50%